FWS parser

EmailValidator/Parser/FoldingWhiteSpace.php

@@ -0,0 +1,92 @@
+<?php
+namespace Egulias\EmailValidator\Parser;
+
+use Egulias\EmailValidator\EmailLexer;
+use Egulias\EmailValidator\Warning\CFWSNearAt;
+use Egulias\EmailValidator\Result\InvalidEmail;
+use Egulias\EmailValidator\Warning\CFWSWithFWS;
+use Egulias\EmailValidator\Result\Reason\CRNoLF;
+use Egulias\EmailValidator\Result\Reason\AtextAfterCFWS;
+use Egulias\EmailValidator\Result\Reason\CRLFAtTheEnd;
+use Egulias\EmailValidator\Result\Reason\CRLFX2;
+use Egulias\EmailValidator\Result\Reason\ExpectingCTEXT;
+use Egulias\EmailValidator\Result\ValidEmail;
+
+class  FoldingWhiteSpace extends Parser
+{
+    public function parse($str)
+    {
+        if (!$this->isFWS()) {
+            return new ValidEmail();
+        }
+
+        $previous = $this->lexer->getPrevious();
+
+        $this->checkCRLFInFWS();
+
+        if ($this->lexer->token['type'] === EmailLexer::S_CR) {
+            return new InvalidEmail(new CRNoLF(), $this->lexer->token['value']);
+        }
+
+        if ($this->lexer->isNextToken(EmailLexer::GENERIC) && $previous['type']  !== EmailLexer::S_AT) {
+            return new InvalidEmail(new AtextAfterCFWS(), $this->lexer->token['value']);
+        }
+
+        if ($this->lexer->token['type'] === EmailLexer::S_LF || $this->lexer->token['type'] === EmailLexer::C_NUL) {
+            return new InvalidEmail(new ExpectingCTEXT(), $this->lexer->token['value']);
+        }
+
+        if ($this->lexer->isNextToken(EmailLexer::S_AT) || $previous['type']  === EmailLexer::S_AT) {
+            $this->warnings[CFWSNearAt::CODE] = new CFWSNearAt();
+        } else {
+            $this->warnings[CFWSWithFWS::CODE] = new CFWSWithFWS();
+        }
+
+        return new ValidEmail();
+    }
+
+    protected function checkCRLFInFWS()
+    {
+        if ($this->lexer->token['type'] !== EmailLexer::CRLF) {
+            return new ValidEmail();
+        }
+
+        if (!$this->lexer->isNextTokenAny(array(EmailLexer::S_SP, EmailLexer::S_HTAB))) {
+            return new InvalidEmail(new CRLFX2(), $this->lexer->token['value']);
+        }
+
+        //this has no coverage. Condition is repeated from above one
+        if (!$this->lexer->isNextTokenAny(array(EmailLexer::S_SP, EmailLexer::S_HTAB))) {
+            return new InvalidEmail(new CRLFAtTheEnd(), $this->lexer->token['value']);
+        }
+    }
+     
+    /**
+     * @return bool
+     */
+    protected function isFWS()
+    {
+        if ($this->escaped()) {
+            return false;
+        }
+
+        return $this->lexer->token['type'] === EmailLexer::S_SP ||
+            $this->lexer->token['type'] === EmailLexer::S_HTAB ||
+            $this->lexer->token['type'] === EmailLexer::S_CR ||
+            $this->lexer->token['type'] === EmailLexer::S_LF ||
+            $this->lexer->token['type'] === EmailLexer::CRLF;
+    }
+
+    /**
+     * @return bool
+     */
+    protected function escaped()
+    {
+        $previous = $this->lexer->getPrevious();
+
+        return $previous && $previous['type'] === EmailLexer::S_BACKSLASH
+            &&
+            $this->lexer->token['type'] !== EmailLexer::GENERIC;
+    }
+
+}

EmailValidator/Parser/LocalPart.php

@@ -26,9 +26,12 @@ class LocalPart extends Parser
             EmailLexer::INVALID => EmailLexer::INVALID
         );
 
+    private $foldingWS;
+
     public function parse($localPart) : Result
     {
         $totalLength = 0;
+        $this->foldingWS = new FoldingWhiteSpace($this->lexer);
 
         while ($this->lexer->token['type'] !== EmailLexer::S_AT && null !== $this->lexer->token['type']) {
             if ($this->hasDotAtStart()) {
@@ -73,8 +76,9 @@ class LocalPart extends Parser
                 return new InvalidEmail(new ReasonExpectingATEXT('Invalid token found'), $this->lexer->token['value']);
             }
 
-            if ($this->isFWS()) {
-                $this->parseFWS();
+            $resultFWS = $this->parseLocalFWS();
+            if($resultFWS->isInvalid()) {
+                return $resultFWS;
             }
 
             $totalLength += strlen($this->lexer->token['value']);
@@ -88,6 +92,18 @@ class LocalPart extends Parser
         return new ValidEmail();
     }
 
+    protected function parseLocalFWS() : Result 
+    {
+        $resultFWS = $this->foldingWS->parse('remove');
+        if ($resultFWS->isValid()) {
+            $warns = $this->foldingWS->getWarnings();
+            foreach ($warns as $code => $dWarning) {
+                $this->warnings[$code] = $dWarning;
+            }
+        }
+        return $resultFWS;
+    }
+
     protected function hasDotAtStart() : bool
     {
             return $this->lexer->token['type'] === EmailLexer::S_DOT && null === $this->lexer->getPrevious()['type'];

EmailValidator/Result/Reason/AtextAfterCFWS.php

@@ -0,0 +1,16 @@
+<?php
+
+namespace Egulias\EmailValidator\Result\Reason;
+
+class AtextAfterCFWS implements Reason
+{
+    public function code() : int
+    {
+        return 133;
+    }
+
+    public function description() : string
+    {
+        return 'ATEXT found after CFWS';
+    }
+}

EmailValidator/Result/Reason/CRLFAtTheEnd.php

@@ -0,0 +1,19 @@
+<?php
+
+namespace Egulias\EmailValidator\Result\Reason;
+
+class CRLFAtTheEnd implements Reason
+{
+    const CODE = 149;
+    const REASON = "CRLF at the end";
+
+    public function code() : int
+    {
+        return 149;
+    }
+
+    public function description() : string
+    {
+        return 'CRLF at the end';
+    }
+}

EmailValidator/Result/Reason/CRLFX2.php

@@ -0,0 +1,16 @@
+<?php
+
+namespace Egulias\EmailValidator\Result\Reason;
+
+class CRLFX2 implements Reason
+{
+    public function code() : int
+    {
+        return 148;
+    }
+
+    public function description() : string
+    {
+        return 'CR  LF tokens found twice';
+    }
+}

EmailValidator/Result/Reason/CRNoLF.php

@@ -0,0 +1,16 @@
+<?php
+
+namespace Egulias\EmailValidator\Result\Reason;
+
+class CRNoLF implements Reason
+{
+    public function code() : int
+    {
+        return 150;
+    }
+
+    public function description() : string
+    {
+        return 'Missing LF after CR';
+    }
+}

EmailValidator/Result/Reason/ExpectingCTEXT.php

@@ -0,0 +1,16 @@
+<?php
+
+namespace Egulias\EmailValidator\Result\Reason;
+
+class ExpectingCTEXT implements Reason
+{
+    public function code() : int
+    {
+        return 139;
+    }
+
+    public function description() : string
+    {
+        return 'Expecting CTEXT';
+    }
+}

Tests/EmailValidator/Validation/RFCValidationTest.php

@@ -36,6 +36,7 @@ use Egulias\EmailValidator\Exception\ExpectingDTEXT;
 use Egulias\EmailValidator\Validation\RFCValidation;
 use Egulias\EmailValidator\Exception\UnclosedComment;
 use Egulias\EmailValidator\Exception\UnopenedComment as ExceptionUnopenedComment;
+use Egulias\EmailValidator\Result\Reason\AtextAfterCFWS as ReasonAtextAfterCFWS;
 use Egulias\EmailValidator\Result\Reason\ConsecutiveDot as ReasonConsecutiveDot;
 use Egulias\EmailValidator\Result\Reason\DotAtEnd as ReasonDotAtEnd;
 use Egulias\EmailValidator\Result\Reason\NoLocalPart;
@@ -44,6 +45,7 @@ use Egulias\EmailValidator\Result\Reason\DotAtStart as ReasonDotAtStart;
 use Egulias\EmailValidator\Result\Reason\ExpectingATEXT as ReasonExpectingATEXT;
 use Egulias\EmailValidator\Result\Reason\UnclosedComment as ReasonUnclosedComment;
 use Egulias\EmailValidator\Result\Reason\UnclosedQuotedString;
+use Egulias\EmailValidator\Result\Reason\CRNoLF as ReasonCRNoLF;
 
 class RFCValidationTest extends TestCase
 {
@@ -217,12 +219,12 @@ class RFCValidationTest extends TestCase
             [new ExceptionUnopenedComment, 'example@(comment))example.com'],
             //This was the original. But atext is not allowed after \n
             //array(EmailValidator::ERR_EXPECTING_ATEXT, "exampl\ne@example.co.uk"),
-            [new AtextAfterCFWS(), "exampl\ne@example.co.uk"],
+            [new InvalidEmail(new ReasonAtextAfterCFWS(), "\n"), "exampl\ne@example.co.uk"],
             [new ExpectingDTEXT(), "example@[[]"],
-            [new AtextAfterCFWS(), "exampl\te@example.co.uk"],
+            [new InvalidEmail(new ReasonAtextAfterCFWS(), "\t"), "exampl\te@example.co.uk"],
             [new CRNoLF(), "example@exa\rmple.co.uk"],
             [new CRNoLF(), "example@[\r]"],
-            [new CRNoLF(), "exam\rple@example.co.uk"],
+            [new InvalidEmail(new ReasonCRNoLF(), "\r"), "exam\rple@example.co.uk"],
         ];
     }